Avnet will promote and resell Mipsology’s Zebra software platform to its APAC customer base. Zebra removes the technical complexity of FPGAs, making them plug-and-play with exceptionally fast performance. This agreement extends Avnet’s IoT ecosystem, bringing Mipsology’s breakthrough deep learning inference acceleration solution to its Asia customers. Companies looking to deploy AI can now seamlessly migrate to new FPGA-based acceleration technologies with no code change and enjoy a much longer lifespan for software and hardware than they could with GPU-based solutions.
Avnet’s first product incorporating the solution will be the Zebra-powered Xilinx Alveo data center accelerator cards. The range of offerings is expected to expand in the future. In addition to distributing the solution, Avnet offers a comprehensive suite of services for customers looking at deploying machine learning for data centers, including hardware, software, system integration, application development, design chain and technical expertise.
FPGAs are better suited than GPUs to accelerate the decision-making (inference) process for large industrial AI applications like video surveillance, smart cities and autonomous cars. However, FPGA programming requires the significant knowledge and expertise of specialized, hard-to-find hardware designers. Zebra eliminates the need for FPGA expertise, making them as easy to use for deep learning inference acceleration as CPU/GPU. Running neural networks defined with TensorFlow, PyTorch, Caffe, and other frameworks on FPGAs would normally require considerable manual time and effort; Zebra makes it instant and effortless.
“Avnet is committed to driving IoT innovation and simplifying complexities, enabling businesses to deploy complete solutions more rapidly,” said Alan Chui, supplier business management president, Avnet Asia. “The combination of Avnet’s comprehensive design services, Alveo’s best-in-class FPGA acceleration and Mipsology’s Zebra technology empowers our customers with a low-cost, high-performance and long lifespan solution for AI neural networking inference.”
